Only 12 out of 50 banks can prove a financial return on AIEvident benchmark study
Why are major banks cutting AI spending?
The computing bill has skyrocketed without delivering the promised financial gains. According to reporting by InfoMoney and data from benchmarking firm Evident, the largest international financial institutions have begun strictly reviewing their technology budgets. The core reason is the absence of clear return-on-investment metrics, which has put chief financial officers on a collision course with innovation departments.
Over recent quarters, banks handed out flexible budgets to avoid falling behind in the tech race. However, following a succession of unexpected cloud infrastructure bills, curbing operational expenses has become an urgent priority in executive suites.
Only 24% of the industry giants tracked by Evident attribute real returns to their projects.
Fell 1.23% during a trading session marked by risk aversion toward growth assets.
AI costs are competing directly with legacy system modernization and cybersecurity.
What caused the tech budget blowout?
Unpredictability in computing token consumption has been the primary budget villain. Financial institutions placed most of their hopes on so-called agentic AI—autonomous systems designed to execute complex workflows without continuous human intervention. However, the more sophisticated the reasoning required of the machine, the more real-time processing power it consumes.
This dynamic prevents CFOs from projecting fixed expenses. The volume of tokens required varies depending on the complexity of each demand, generating open-ended bills that disrupt cost planning at large corporations.
“In truth, I don’t think anyone inside the organization has a good handle on this—if I’m going to do XYZ, what is my cost going to be? As AI advances toward more sophisticated reasoning and autonomous workflows, the tokens required for a given task can vary significantly.” Bhavi Mehta, partner at Bain & Co. and global head of advanced analytics in financial services
As Bhavi Mehta pointed out, understanding costs relative to the actual value delivered to the business has become essential. Banks are now demanding proof of efficiency before authorizing any scaling up of models.
Which expenses are competing with AI at financial institutions?
Artificial intelligence has begun competing for resources with mandatory maintenance demands. As reported by InfoMoney, banks operate under constant pressure to modernize legacy systems—older frameworks that still support core payment and account processing—and to reinforce cybersecurity defenses against corporate attacks.
While digital security and legacy system maintenance prevent immediate operational collapses, AI had been operating as an experimental bet. The moment cloud computing costs explode without clear proof of additional revenue, executive leadership cuts spending first where the return is hypothetical.
Warning for the corporate market: Only 12 of the 50 largest global banks tracked by Evident are able to disclose return projections or concrete metrics tied to their AI spending. The remaining 38 have yet to financially demonstrate the effectiveness of their outlays.
How is banking caution impacting the S&P 500 and tech companies?
The retreat by banks directly affects projected revenues for AI infrastructure providers. This shift in international banking balance sheets reinforces a defensive posture toward U.S.-listed tech companies whose stock market multiples factor in assumptions of continuous, accelerated monetization.
The Dow Jones Industrial Average dropped 1.23% on the day, closing at 52,215.16 points, illustrating broader caution on Wall Street. If major institutional buyers—such as banks—shut off the taps to audit projects, expected revenue for software providers, cloud platforms, and model developers is likely to slow down.
Investors holding ETFs tied to the S&P 500 or technology BDRs need to pay closer attention to valuation multiples. These companies' share prices were pricing in corporate clients maintaining growing outlays without questioning margins. With banks demanding productivity metrics before renewing contracts, the sector's growth pace may be marked down.
What should investors monitor from here on out?
The market is entering a phase of reckoning for actual operational delivery. The period of blank-check testing is behind us, making way for rigorous audits of efficiency and profit delivery capacity.
Disclosure of return metrics — Monitor whether the number of banks with proven ROI according to Evident goes beyond the current 12 institutions in upcoming earnings rounds.
Cloud consumption contracts — Observe whether quarterly results from major infrastructure providers report renegotiations toward fixed pricing models rather than open-ended token consumption.
Multiple compression in the S&P 500 — Track whether the brake on financial spending forces analysts to recalibrate revenue estimates for companies whose stocks rose on AI euphoria.
Rico aos Poucos Verdict: The banks' pullback does not spell the end of the technology, but it marks the close of the uncontrolled experimentation phase. For investors, the current moment requires separating theses anchored in real results from those that relied on corporate blank checks to justify stretched valuations.
